Research in three areas is proposed in this project: 1) Optimal design of clinical trials for comparing two or more treatments, and applications of sequential methods to resolve certain basic ethical problems involved. 2) Adaptive design in regression for the determination of the optimal dosage levels of treatments to secure a desired mean response, and applications of stochastic approximation methods to these and other problems in the control of biological systems. 3) Empirical Bayes methods in studying mean levels and variability, and applications of these methods to clinical problems.